A UAV Target Recognition Method Based on Micro-Doppler Feature Fusion

This study introduces a micro-Doppler-centric approach to UAV recognition, aiming to overcome the limitations associated with a single feature. A multi-interaction multi-feature fusion framework is proposed, encompassing four types of acquired signals, which undergo preprocessing. Subsequently, graphical features are extracted using a designated backbone network. The final step involves the fusion of these four distinct graphical features to enable UAV identification by the deep neural network. Empirical results substantiate the effectiveness of feature fusion, demonstrating superior accuracy in UAV recognition compared to relying solely on individual signal image features.